Subject: Re: IBM Workpad z50 Vertical Bars Problem
To: Florian Cramer <cantsin@zedat.fu-berlin.de>
From: Bryan Vyhmeister <bsd@hub3.net>
List: port-hpcmips
Date: 06/11/2003 22:42:05
On Tuesday, June 10, 2003, at 12:04 PM, Florian Cramer wrote:

> Am Montag, 09. Juni 2003 um 02:48:47 Uhr (-0700) schrieb Siva Arumugam:
>> I haven't had a vertical bars problem in quite a
>> while. I used to be sure (!) that it's because I
>> always "sync".
>
> Hi, I probably was the one who first report "sync" as a workaround for
> the problem. It turns out, at least on my z50, that this
> increases the probability to resume the unit into a working state, but
> doesn't solve the issue as such. Now, I get about 10% bad resumes with
> vertical bars.

Hmm. Interesting.

> - So I guess it's a problem with files being open/written during the
> suspend cycle?

Kind of sounds that way. The suspend function is a function of the 
hardware, but it is supported by NetBSD also, correct? I know that on 
some laptops, the suspend function seems to almost be totally done in 
hardware including a keystroke to activate the suspend. Others seem 
very "software" connected. Would there be any way to add some kernel 
code to sync each time a suspend is detected? It could be an option 
like Z50_SYNC_SUSPEND or something. I guess this is not practical 
unless it is for sure files being open or written that is causing the 
problem.

Is there any difference if you are using a PCMCIA memory card or hard 
drive versus a Compact Flash memory card or microdrive? That's 
something I plan to test in July after I get back from a two week trip.

Bryan